Copyright of Informa UK Ltd. Printing and distribution strictly prohibited Review 10.1517/17425250802115025 © 2008 Informa UK Ltd ISSN 1742-5255 733 High-throughput enzymology and combinatorial mutagenesis for mining cytochrome P450 functions Philippe Urban, Gilles Truan & Denis Pompon † Centre de Génétique Moléculaire, Laboratoire d’Ingénierie des Protéines Membranaires, Gif-sur-Yvette, France Background: High-throughput (HT) characterization of drugs for potential biotransformation and interaction is routine in pharmaceutical industry. Objective: HT approaches were extended to enzyme studies for identifying combinations of structural elements that control substrate specificity. Methods: Structure-based and combinatorial mutagenesis have been applied with success to decipher P450 structure–function relationships. The idea is to measure activities on a library of combinatorial variants of similar structure with a large collection of substrates presenting a similar chemical scaffold. This combinatorial approach is then associated to multivariate statistics to relate functional features to structural determinants. Results/conclusion: A method to measure HT kinetics is presented. The proposed statistical approach is illustrated with tri- and tetracyclic substrates and artificial variant enzymes of the CYP1A subfamily. Keywords: combinatorial mutagenesis, cytochromes P450, drug development, functional prediction, multivariate analysis, structure–activity relationships Expert Opin. Drug Metab. Toxicol. (2008) 4(6):733-747 1. Introduction Defining a set of objective criteria to describe substrate specificity is of key importance for structure–function relationship studies and bioengineering [1]. Enzyme substrate specificity can be viewed as the limited collection of molecules that bind to and are transformed by a given enzyme. In a classical view, most enzymes transform a single substrate that specifies the enzyme by giving it its name. This high specificity was first explained by Fisher as the template hypothesis [2], and then extended in terms of both chemistry and protein flexibility by Koshland [3]. It is the consequence of the necessary fitness of biological processes within the cell, where each biosynthetic step is associated with a specialized enzyme [4]. However, the notion of substrate specificity is blurred; even more today than was the case some decades ago. The development of biosynthetic pathway engineering and of high-throughput (HT) activity screening technologies revealed that other activities unrelated to the main activity can be catalyzed, even for enzymes with narrow substrate specificity. For instance, human carbonic anhydrase forms bicarbonate by catalyzing reversible hydration of carbon dioxide [5]. But despite a well-defined natural substrate, carbonic anhydrase can also catalyze an esterase activity toward the unrelated substrate 2-naphthyl acetate [6], which would have been difficult to predict on the basis of an analogy to carbon dioxide. Such latent, promiscuous activities are frequently significantly slower than main activity [7]. A number of enzymes can act with a similar efficiency on several alternate substrates, and should therefore be classified as 1. Introduction 2. Cytochromes P450 with fuzzy substrate specificity 3. Combinatorial approach versus site-directed mutagenesis 4. Experimental and computational models 5. Combinatorial libraries of P450 enzymes and of substrates 6. High-throughput activity screening 7. High-throughput generated structure–activity matrices 8. Multivariate analysis of structure–activity matrices 9. Recent studies with CYP1A enzymes 10. Expert opinion